Entries Tagged "artificial intelligence"

Page 3 of 8

Class-Action Lawsuit for Scraping Data without Permission

I have mixed feelings about this class-action lawsuit against OpenAI and Microsoft, claiming that it “scraped 300 billion words from the internet” without either registering as a data broker or obtaining consent. On the one hand, I want this to be a protected fair use of public data. On the other hand, I want us all to be compensated for our uniquely human ability to generate language.

There’s an interesting wrinkle on this. A recent paper showed that using AI generated text to train another AI invariably “causes irreversible defects.” From a summary:

The tails of the original content distribution disappear. Within a few generations, text becomes garbage, as Gaussian distributions converge and may even become delta functions. We call this effect model collapse.

Just as we’ve strewn the oceans with plastic trash and filled the atmosphere with carbon dioxide, so we’re about to fill the Internet with blah. This will make it harder to train newer models by scraping the web, giving an advantage to firms which already did that, or which control access to human interfaces at scale. Indeed, we already see AI startups hammering the Internet Archive for training data.

This is the same idea that Ted Chiang wrote about: that ChatGPT is a “blurry JPEG of all the text on the Web.” But the paper includes the math that proves the claim.

What this means is that text from before last year—text that is known human-generated—will become increasingly valuable.

Posted on July 5, 2023 at 7:14 AMView Comments

AI as Sensemaking for Public Comments

It’s become fashionable to think of artificial intelligence as an inherently dehumanizing technology, a ruthless force of automation that has unleashed legions of virtual skilled laborers in faceless form. But what if AI turns out to be the one tool able to identify what makes your ideas special, recognizing your unique perspective and potential on the issues where it matters most?

You’d be forgiven if you’re distraught about society’s ability to grapple with this new technology. So far, there’s no lack of prognostications about the democratic doom that AI may wreak on the US system of government. There are legitimate reasons to be concerned that AI could spread misinformation, break public comment processes on regulations, inundate legislators with artificial constituent outreach, help to automate corporate lobbying, or even generate laws in a way tailored to benefit narrow interests.

But there are reasons to feel more sanguine as well. Many groups have started demonstrating the potential beneficial uses of AI for governance. A key constructive-use case for AI in democratic processes is to serve as discussion moderator and consensus builder.

To help democracy scale better in the face of growing, increasingly interconnected populations—as well as the wide availability of AI language tools that can generate reams of text at the click of a button—the US will need to leverage AI’s capability to rapidly digest, interpret and summarize this content.

There are two different ways to approach the use of generative AI to improve civic participation and governance. Each is likely to lead to drastically different experience for public policy advocates and other people trying to have their voice heard in a future system where AI chatbots are both the dominant readers and writers of public comment.

For example, consider individual letters to a representative, or comments as part of a regulatory rulemaking process. In both cases, we the people are telling the government what we think and want.

For more than half a century, agencies have been using human power to read through all the comments received, and to generate summaries and responses of their major themes. To be sure, digital technology has helped.

In 2021, the Council of Federal Chief Data Officers recommended modernizing the comment review process by implementing natural language processing tools for removing duplicates and clustering similar comments in processes governmentwide. These tools are simplistic by the standards of 2023 AI. They work by assessing the semantic similarity of comments based on metrics like word frequency (How often did you say “personhood”?) and clustering similar comments and giving reviewers a sense of what topic they relate to.

Think of this approach as collapsing public opinion. They take a big, hairy mass of comments from thousands of people and condense them into a tidy set of essential reading that generally suffices to represent the broad themes of community feedback. This is far easier for a small agency staff or legislative office to handle than it would be for staffers to actually read through that many individual perspectives.

But what’s lost in this collapsing is individuality, personality, and relationships. The reviewer of the condensed comments may miss the personal circumstances that led so many commenters to write in with a common point of view, and may overlook the arguments and anecdotes that might be the most persuasive content of the testimony.

Most importantly, the reviewers may miss out on the opportunity to recognize committed and knowledgeable advocates, whether interest groups or individuals, who could have long-term, productive relationships with the agency.

These drawbacks have real ramifications for the potential efficacy of those thousands of individual messages, undermining what all those people were doing it for. Still, practicality tips the balance toward of some kind of summarization approach. A passionate letter of advocacy doesn’t hold any value if regulators or legislators simply don’t have time to read it.

There is another approach. In addition to collapsing testimony through summarization, government staff can use modern AI techniques to explode it. They can automatically recover and recognize a distinctive argument from one piece of testimony that does not exist in the thousands of other testimonies received. They can discover the kinds of constituent stories and experiences that legislators love to repeat at hearings, town halls and campaign events. This approach can sustain the potential impact of individual public comment to shape legislation even as the volumes of testimony may rise exponentially.

In computing, there is a rich history of that type of automation task in what is called outlier detection. Traditional methods generally involve finding a simple model that explains most of the data in question, like a set of topics that well describe the vast majority of submitted comments. But then they go a step further by isolating those data points that fall outside the mold—comments that don’t use arguments that fit into the neat little clusters.

State-of-the-art AI language models aren’t necessary for identifying outliers in text document data sets, but using them could bring a greater degree of sophistication and flexibility to this procedure. AI language models can be tasked to identify novel perspectives within a large body of text through prompting alone. You simply need to tell the AI to find them.

In the absence of that ability to extract distinctive comments, lawmakers and regulators have no choice but to prioritize on other factors. If there is nothing better, “who donated the most to our campaign” or “which company employs the most of my former staffers” become reasonable metrics for prioritizing public comments. AI can help elected representatives do much better.

If Americans want AI to help revitalize the country’s ailing democracy, they need to think about how to align the incentives of elected leaders with those of individuals. Right now, as much as 90% of constituent communications are mass emails organized by advocacy groups, and they go largely ignored by staffers. People are channeling their passions into a vast digital warehouses where algorithms box up their expressions so they don’t have to be read. As a result, the incentive for citizens and advocacy groups is to fill that box up to the brim, so someone will notice it’s overflowing.

A talented, knowledgeable, engaged citizen should be able to articulate their ideas and share their personal experiences and distinctive points of view in a way that they can be both included with everyone else’s comments where they contribute to summarization and recognized individually among the other comments. An effective comment summarization process would extricate those unique points of view from the pile and put them into lawmakers’ hands.

This essay was written with Nathan Sanders, and previously appeared in the Conversation.

Posted on June 22, 2023 at 11:43 AMView Comments

On the Need for an AI Public Option

Artificial intelligence will bring great benefits to all of humanity. But do we really want to entrust this revolutionary technology solely to a small group of US tech companies?

Silicon Valley has produced no small number of moral disappointments. Google retired its “don’t be evil” pledge before firing its star ethicist. Self-proclaimed “free speech absolutist” Elon Musk bought Twitter in order to censor political speech, retaliate against journalists, and ease access to the platform for Russian and Chinese propagandists. Facebook lied about how it enabled Russian interference in the 2016 US presidential election and paid a public relations firm to blame Google and George Soros instead.

These and countless other ethical lapses should prompt us to consider whether we want to give technology companies further abilities to learn our personal details and influence our day-to-day decisions. Tech companies can already access our daily whereabouts and search queries. Digital devices monitor more and more aspects of our lives: We have cameras in our homes and heartbeat sensors on our wrists sending what they detect to Silicon Valley.

Now, tech giants are developing ever more powerful AI systems that don’t merely monitor you; they actually interact with you—and with others on your behalf. If searching on Google in the 2010s was like being watched on a security camera, then using AI in the late 2020s will be like having a butler. You will willingly include them in every conversation you have, everything you write, every item you shop for, every want, every fear, everything. It will never forget. And, despite your reliance on it, it will be surreptitiously working to further the interests of one of these for-profit corporations.

There’s a reason Google, Microsoft, Facebook, and other large tech companies are leading the AI revolution: Building a competitive large language model (LLM) like the one powering ChatGPT is incredibly expensive. It requires upward of $100 million in computational costs for a single model training run, in addition to access to large amounts of data. It also requires technical expertise, which, while increasingly open and available, remains heavily concentrated in a small handful of companies. Efforts to disrupt the AI oligopoly by funding start-ups are self-defeating as Big Tech profits from the cloud computing services and AI models powering those start-ups—and often ends up acquiring the start-ups themselves.

Yet corporations aren’t the only entities large enough to absorb the cost of large-scale model training. Governments can do it, too. It’s time to start taking AI development out of the exclusive hands of private companies and bringing it into the public sector. The United States needs a government-funded-and-directed AI program to develop widely reusable models in the public interest, guided by technical expertise housed in federal agencies.

So far, the AI regulation debate in Washington has focused on the governance of private-sector activity—which the US Congress is in no hurry to advance. Congress should not only hurry up and push AI regulation forward but also go one step further and develop its own programs for AI. Legislators should reframe the AI debate from one about public regulation to one about public development.

The AI development program could be responsive to public input and subject to political oversight. It could be directed to respond to critical issues such as privacy protection, underpaid tech workers, AI’s horrendous carbon emissions, and the exploitation of unlicensed data. Compared to keeping AI in the hands of morally dubious tech companies, the public alternative is better both ethically and economically. And the switch should take place soon: By the time AI becomes critical infrastructure, essential to large swaths of economic activity and daily life, it will be too late to get started.

Other countries are already there. China has heavily prioritized public investment in AI research and development by betting on a handpicked set of giant companies that are ostensibly private but widely understood to be an extension of the state. The government has tasked Alibaba, Huawei, and others with creating products that support the larger ecosystem of state surveillance and authoritarianism.

The European Union is also aggressively pushing AI development. The European Commission already invests 1 billion euros per year in AI, with a plan to increase that figure to 20 billion euros annually by 2030. The money goes to a continent-wide network of public research labs, universities, and private companies jointly working on various parts of AI. The Europeans’ focus is on knowledge transfer, developing the technology sector, use of AI in public administration, mitigating safety risks, and preserving fundamental rights. The EU also continues to be at the cutting edge of aggressively regulating both data and AI.

Neither the Chinese nor the European model is necessarily right for the United States. State control of private enterprise remains anathema in American political culture and would struggle to gain mainstream traction. The tech companies—and their supporters in both US political parties—are opposed to robust public governance of AI. But Washington can take inspiration from China and Europe’;s long-range planning and leadership on regulation and public investment. With boosters pointing to hundreds of trillions of dollars of global economic value associated with AI, the stakes of international competition are compelling. As in energy and medical research, which have their own federal agencies in the Department of Energy and the National Institutes of Health, respectively, there is a place for AI research and development inside government.

Beside the moral argument against letting private companies develop AI, there’s a strong economic argument in favor of a public option as well. A publicly funded LLM could serve as an open platform for innovation, helping any small business, nonprofit, or individual entrepreneur to build AI-assisted applications.

There’s also a practical argument. Building AI is within public reach because governments don’t need to own and operate the entire AI supply chain. Chip and computer production, cloud data centers, and various value-added applications—such as those that integrate AI with consumer electronics devices or entertainment software—do not need to be publicly controlled or funded.

One reason to be skeptical of public funding for AI is that it might result in a lower quality and slower innovation, given greater ethical scrutiny, political constraints, and fewer incentives due to a lack of market competition. But even if that is the case, it would be worth broader access to the most important technology of the 21st century. And it is by no means certain that public AI has to be at a disadvantage. The open-source community is proof that it’s not always private companies that are the most innovative.

Those who worry about the quality trade-off might suggest a public buyer model, whereby Washington licenses or buys private language models from Big Tech instead of developing them itself. But that doesn’t go far enough to ensure that the tools are aligned with public priorities and responsive to public needs. It would not give the public detailed insight into or control of the inner workings and training procedures for these models, and it would still require strict and complex regulation.

There is political will to take action to develop AI via public, rather than private, funds—but this does not yet equate to the will to create a fully public AI development agency. A task force created by Congress recommended in January a $2.6 billion federal investment in computing and data resources to prime the AI research ecosystem in the United States. But this investment would largely serve to advance the interests of Big Tech, leaving the opportunity for public ownership and oversight unaddressed.

Nonprofit and academic organizations have already created open-access LLMs. While these should be celebrated, they are not a substitute for a public option. Nonprofit projects are still beholden to private interests, even if they are benevolent ones. These private interests can change without public input, as when OpenAI effectively abandoned its nonprofit origins, and we can’t be sure that their founding intentions or operations will survive market pressures, fickle donors, and changes in leadership.

The US government is by no means a perfect beacon of transparency, a secure and responsible store of our data, or a genuine reflection of the public’s interests. But the risks of placing AI development entirely in the hands of demonstrably untrustworthy Silicon Valley companies are too high. AI will impact the public like few other technologies, so it should also be developed by the public.

This essay was written with Nathan Sanders, and appeared in Foreign Policy.

Posted on June 14, 2023 at 7:02 AMView Comments

AI-Generated Steganography

New research suggests that AIs can produce perfectly secure steganographic images:

Abstract: Steganography is the practice of encoding secret information into innocuous content in such a manner that an adversarial third party would not realize that there is hidden meaning. While this problem has classically been studied in security literature, recent advances in generative models have led to a shared interest among security and machine learning researchers in developing scalable steganography techniques. In this work, we show that a steganography procedure is perfectly secure under Cachin (1998)’s information theoretic-model of steganography if and only if it is induced by a coupling. Furthermore, we show that, among perfectly secure procedures, a procedure is maximally efficient if and only if it is induced by a minimum entropy coupling. These insights yield what are, to the best of our knowledge, the first steganography algorithms to achieve perfect security guarantees with non-trivial efficiency; additionally, these algorithms are highly scalable. To provide empirical validation, we compare a minimum entropy coupling-based approach to three modern baselines—arithmetic coding, Meteor, and adaptive dynamic grouping—using GPT-2, WaveRNN, and Image Transformer as communication channels. We find that the minimum entropy coupling-based approach achieves superior encoding efficiency, despite its stronger security constraints. In aggregate, these results suggest that it may be natural to view information-theoretic steganography through the lens of minimum entropy coupling.

News article.

EDITED TO ADD (6/13): Comments.

Posted on June 12, 2023 at 7:18 AMView Comments

Open-Source LLMs

In February, Meta released its large language model: LLaMA. Unlike OpenAI and its ChatGPT, Meta didn’t just give the world a chat window to play with. Instead, it released the code into the open-source community, and shortly thereafter the model itself was leaked. Researchers and programmers immediately started modifying it, improving it, and getting it to do things no one else anticipated. And their results have been immediate, innovative, and an indication of how the future of this technology is going to play out. Training speeds have hugely increased, and the size of the models themselves has shrunk to the point that you can create and run them on a laptop. The world of AI research has dramatically changed.

This development hasn’t made the same splash as other corporate announcements, but its effects will be much greater. It will wrest power from the large tech corporations, resulting in both much more innovation and a much more challenging regulatory landscape. The large corporations that had controlled these models warn that this free-for-all will lead to potentially dangerous developments, and problematic uses of the open technology have already been documented. But those who are working on the open models counter that a more democratic research environment is better than having this powerful technology controlled by a small number of corporations.

The power shift comes from simplification. The LLMs built by OpenAI and Google rely on massive data sets, measured in the tens of billions of bytes, computed on by tens of thousands of powerful specialized processors producing models with billions of parameters. The received wisdom is that bigger data, bigger processing, and larger parameter sets were all needed to make a better model. Producing such a model requires the resources of a corporation with the money and computing power of a Google or Microsoft or Meta.

But building on public models like Meta’s LLaMa, the open-source community has innovated in ways that allow results nearly as good as the huge models—but run on home machines with common data sets. What was once the reserve of the resource-rich has become a playground for anyone with curiosity, coding skills, and a good laptop. Bigger may be better, but the open-source community is showing that smaller is often good enough. This opens the door to more efficient, accessible, and resource-friendly LLMs.

More importantly, these smaller and faster LLMs are much more accessible and easier to experiment with. Rather than needing tens of thousands of machines and millions of dollars to train a new model, an existing model can now be customized on a mid-priced laptop in a few hours. This fosters rapid innovation.

It also takes control away from large companies like Google and OpenAI. By providing access to the underlying code and encouraging collaboration, open-source initiatives empower a diverse range of developers, researchers, and organizations to shape the technology. This diversification of control helps prevent undue influence, and ensures that the development and deployment of AI technologies align with a broader set of values and priorities. Much of the modern internet was built on open-source technologies from the LAMP (Linux, Apache, mySQL, and PHP/PERL/Python) stack—a suite of applications often used in web development. This enabled sophisticated websites to be easily constructed, all with open-source tools that were built by enthusiasts, not companies looking for profit. Facebook itself was originally built using open-source PHP.

But being open-source also means that there is no one to hold responsible for misuse of the technology. When vulnerabilities are discovered in obscure bits of open-source technology critical to the functioning of the internet, often there is no entity responsible for fixing the bug. Open-source communities span countries and cultures, making it difficult to ensure that any country’s laws will be respected by the community. And having the technology open-sourced means that those who wish to use it for unintended, illegal, or nefarious purposes have the same access to the technology as anyone else.

This, in turn, has significant implications for those who are looking to regulate this new and powerful technology. Now that the open-source community is remixing LLMs, it’s no longer possible to regulate the technology by dictating what research and development can be done; there are simply too many researchers doing too many different things in too many different countries. The only governance mechanism available to governments now is to regulate usage (and only for those who pay attention to the law), or to offer incentives to those (including startups, individuals, and small companies) who are now the drivers of innovation in the arena. Incentives for these communities could take the form of rewards for the production of particular uses of the technology, or hackathons to develop particularly useful applications. Sticks are hard to use—instead, we need appealing carrots.

It is important to remember that the open-source community is not always motivated by profit. The members of this community are often driven by curiosity, the desire to experiment, or the simple joys of building. While there are companies that profit from supporting software produced by open-source projects like Linux, Python, or the Apache web server, those communities are not profit driven.

And there are many open-source models to choose from. Alpaca, Cerebras-GPT, Dolly, HuggingChat, and StableLM have all been released in the past few months. Most of them are built on top of LLaMA, but some have other pedigrees. More are on their way.

The large tech monopolies that have been developing and fielding LLMs—Google, Microsoft, and Meta—are not ready for this. A few weeks ago, a Google employee leaked a memo in which an engineer tried to explain to his superiors what an open-source LLM means for their own proprietary tech. The memo concluded that the open-source community has lapped the major corporations and has an overwhelming lead on them.

This isn’t the first time companies have ignored the power of the open-source community. Sun never understood Linux. Netscape never understood the Apache web server. Open source isn’t very good at original innovations, but once an innovation is seen and picked up, the community can be a pretty overwhelming thing. The large companies may respond by trying to retrench and pulling their models back from the open-source community.

But it’s too late. We have entered an era of LLM democratization. By showing that smaller models can be highly effective, enabling easy experimentation, diversifying control, and providing incentives that are not profit motivated, open-source initiatives are moving us into a more dynamic and inclusive AI landscape. This doesn’t mean that some of these models won’t be biased, or wrong, or used to generate disinformation or abuse. But it does mean that controlling this technology is going to take an entirely different approach than regulating the large players.

This essay was written with Jim Waldo, and previously appeared on Slate.com.

EDITED TO ADD (6/4): Slashdot thread.

Posted on June 2, 2023 at 10:21 AMView Comments

On the Catastrophic Risk of AI

Earlier this week, I signed on to a short group statement, coordinated by the Center for AI Safety:

Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.

The press coverage has been extensive, and surprising to me. The New York Times headline is “A.I. Poses ‘Risk of Extinction,’ Industry Leaders Warn.” BBC: “Artificial intelligence could lead to extinction, experts warn.” Other headlines are similar.

I actually don’t think that AI poses a risk to human extinction. I think it poses a similar risk to pandemics and nuclear war—which is to say, a risk worth taking seriously, but not something to panic over. Which is what I thought the statement said.

In my talk at the RSA Conference last month, I talked about the power level of our species becoming too great for our systems of governance. Talking about those systems, I said:

Now, add into this mix the risks that arise from new and dangerous technologies such as the internet or AI or synthetic biology. Or molecular nanotechnology, or nuclear weapons. Here, misaligned incentives and hacking can have catastrophic consequences for society.

That was what I was thinking about when I agreed to sign on to the statement: “Pandemics, nuclear weapons, AI—yeah, I would put those three in the same bucket. Surely we can spend the same effort on AI risk as we do on future pandemics. That’s a really low bar.” Clearly I should have focused on the word “extinction,” and not the relative comparisons.

Seth Lazar, Jeremy Howard, and Arvind Narayanan wrote:

We think that, in fact, most signatories to the statement believe that runaway AI is a way off yet, and that it will take a significant scientific advance to get there­—ne that we cannot anticipate, even if we are confident that it will someday occur. If this is so, then at least two things follow.

I agree with that, and with their follow up:

First, we should give more weight to serious risks from AI that are more urgent. Even if existing AI systems and their plausible extensions won’t wipe us out, they are already causing much more concentrated harm, they are sure to exacerbate inequality and, in the hands of power-hungry governments and unscrupulous corporations, will undermine individual and collective freedom.

This is what I wrote in Click Here to Kill Everybody (2018):

I am less worried about AI; I regard fear of AI more as a mirror of our own society than as a harbinger of the future. AI and intelligent robotics are the culmination of several precursor technologies, like machine learning algorithms, automation, and autonomy. The security risks from those precursor technologies are already with us, and they’re increasing as the technologies become more powerful and more prevalent. So, while I am worried about intelligent and even driverless cars, most of the risks arealready prevalent in Internet-connected drivered cars. And while I am worried about robot soldiers, most of the risks are already prevalent in autonomous weapons systems.

Also, as roboticist Rodney Brooks pointed out, “Long before we see such machines arising there will be the somewhat less intelligent and belligerent machines. Before that there will be the really grumpy machines. Before that the quite annoying machines. And before them the arrogant unpleasant machines.” I think we’ll see any new security risks coming long before they get here.

I do think we should worry about catastrophic AI and robotics risk. It’s the fact that they affect the world in a direct, physical manner—and that they’re vulnerable to class breaks.

(Other things to read: David Chapman is good on scary AI. And Kieran Healy is good on the statement.)

Okay, enough. I should also learn not to sign on to group statements.

EDITED TO ADD (9/9): The Brooks quote is from this excellent essay.

Posted on June 1, 2023 at 7:17 AMView Comments

On the Poisoning of LLMs

Interesting essay on the poisoning of LLMs—ChatGPT in particular:

Given that we’ve known about model poisoning for years, and given the strong incentives the black-hat SEO crowd has to manipulate results, it’s entirely possible that bad actors have been poisoning ChatGPT for months. We don’t know because OpenAI doesn’t talk about their processes, how they validate the prompts they use for training, how they vet their training data set, or how they fine-tune ChatGPT. Their secrecy means we don’t know if ChatGPT has been safely managed.

They’ll also have to update their training data set at some point. They can’t leave their models stuck in 2021 forever.

Once they do update it, we only have their word—pinky-swear promises—that they’ve done a good enough job of filtering out keyword manipulations and other training data attacks, something that the AI researcher El Mahdi El Mhamdi posited is mathematically impossible in a paper he worked on while he was at Google.

Posted on May 25, 2023 at 7:05 AMView Comments

Credible Handwriting Machine

In case you don’t have enough to worry about, someone has built a credible handwriting machine:

This is still a work in progress, but the project seeks to solve one of the biggest problems with other homework machines, such as this one that I covered a few months ago after it blew up on social media. The problem with most homework machines is that they’re too perfect. Not only is their content output too well-written for most students, but they also have perfect grammar and punctuation ­ something even we professional writers fail to consistently achieve. Most importantly, the machine’s “handwriting” is too consistent. Humans always include small variations in their writing, no matter how honed their penmanship.

Devadath is on a quest to fix the issue with perfect penmanship by making his machine mimic human handwriting. Even better, it will reflect the handwriting of its specific user so that AI-written submissions match those written by the student themselves.

Like other machines, this starts with asking ChatGPT to write an essay based on the assignment prompt. That generates a chunk of text, which would normally be stylized with a script-style font and then output as g-code for a pen plotter. But instead, Devadeth created custom software that records examples of the user’s own handwriting. The software then uses that as a font, with small random variations, to create a document image that looks like it was actually handwritten.

Watch the video.

My guess is that this is another detection/detection avoidance arms race.

Posted on May 23, 2023 at 7:15 AMView Comments

Ted Chiang on the Risks of AI

Ted Chiang has an excellent essay in the New Yorker: “Will A.I. Become the New McKinsey?”

The question we should be asking is: as A.I. becomes more powerful and flexible, is there any way to keep it from being another version of McKinsey? The question is worth considering across different meanings of the term “A.I.” If you think of A.I. as a broad set of technologies being marketed to companies to help them cut their costs, the question becomes: how do we keep those technologies from working as “capital’s willing executioners”? Alternatively, if you imagine A.I. as a semi-autonomous software program that solves problems that humans ask it to solve, the question is then: how do we prevent that software from assisting corporations in ways that make people’s lives worse? Suppose you’ve built a semi-autonomous A.I. that’s entirely obedient to humans­—one that repeatedly checks to make sure it hasn’t misinterpreted the instructions it has received. This is the dream of many A.I. researchers. Yet such software could easily still cause as much harm as McKinsey has.

Note that you cannot simply say that you will build A.I. that only offers pro-social solutions to the problems you ask it to solve. That’s the equivalent of saying that you can defuse the threat of McKinsey by starting a consulting firm that only offers such solutions. The reality is that Fortune 100 companies will hire McKinsey instead of your pro-social firm, because McKinsey’s solutions will increase shareholder value more than your firm’s solutions will. It will always be possible to build A.I. that pursues shareholder value above all else, and most companies will prefer to use that A.I. instead of one constrained by your principles.

EDITED TO ADD: Ted Chiang’s previous essay, “ChatGPT Is a Blurry JPEG of the Web” is also worth reading.

Posted on May 12, 2023 at 10:00 AMView Comments

Building Trustworthy AI

We will all soon get into the habit of using AI tools for help with everyday problems and tasks. We should get in the habit of questioning the motives, incentives, and capabilities behind them, too.

Imagine you’re using an AI chatbot to plan a vacation. Did it suggest a particular resort because it knows your preferences, or because the company is getting a kickback from the hotel chain? Later, when you’re using another AI chatbot to learn about a complex economic issue, is the chatbot reflecting your politics or the politics of the company that trained it?

For AI to truly be our assistant, it needs to be trustworthy. For it to be trustworthy, it must be under our control; it can’t be working behind the scenes for some tech monopoly. This means, at a minimum, the technology needs to be transparent. And we all need to understand how it works, at least a little bit.

Amid the myriad warnings about creepy risks to well-being, threats to democracy, and even existential doom that have accompanied stunning recent developments in artificial intelligence (AI)—and large language models (LLMs) like ChatGPT and GPT-4—one optimistic vision is abundantly clear: this technology is useful. It can help you find information, express your thoughts, correct errors in your writing, and much more. If we can navigate the pitfalls, its assistive benefit to humanity could be epoch-defining. But we’re not there yet.

Let’s pause for a moment and imagine the possibilities of a trusted AI assistant. It could write the first draft of anything: emails, reports, essays, even wedding vows. You would have to give it background information and edit its output, of course, but that draft would be written by a model trained on your personal beliefs, knowledge, and style. It could act as your tutor, answering questions interactively on topics you want to learn about—in the manner that suits you best and taking into account what you already know. It could assist you in planning, organizing, and communicating: again, based on your personal preferences. It could advocate on your behalf with third parties: either other humans or other bots. And it could moderate conversations on social media for you, flagging misinformation, removing hate or trolling, translating for speakers of different languages, and keeping discussions on topic; or even mediate conversations in physical spaces, interacting through speech recognition and synthesis capabilities.

Today’s AIs aren’t up for the task. The problem isn’t the technology—that’s advancing faster than even the experts had guessed—it’s who owns it. Today’s AIs are primarily created and run by large technology companies, for their benefit and profit. Sometimes we are permitted to interact with the chatbots, but they’re never truly ours. That’s a conflict of interest, and one that destroys trust.

The transition from awe and eager utilization to suspicion to disillusionment is a well worn one in the technology sector. Twenty years ago, Google’s search engine rapidly rose to monopolistic dominance because of its transformative information retrieval capability. Over time, the company’s dependence on revenue from search advertising led them to degrade that capability. Today, many observers look forward to the death of the search paradigm entirely. Amazon has walked the same path, from honest marketplace to one riddled with lousy products whose vendors have paid to have the company show them to you. We can do better than this. If each of us are going to have an AI assistant helping us with essential activities daily and even advocating on our behalf, we each need to know that it has our interests in mind. Building trustworthy AI will require systemic change.

First, a trustworthy AI system must be controllable by the user. That means that the model should be able to run on a user’s owned electronic devices (perhaps in a simplified form) or within a cloud service that they control. It should show the user how it responds to them, such as when it makes queries to search the web or external services, when it directs other software to do things like sending an email on a user’s behalf, or modifies the user’s prompts to better express what the company that made it thinks the user wants. It should be able to explain its reasoning to users and cite its sources. These requirements are all well within the technical capabilities of AI systems.

Furthermore, users should be in control of the data used to train and fine-tune the AI system. When modern LLMs are built, they are first trained on massive, generic corpora of textual data typically sourced from across the Internet. Many systems go a step further by fine-tuning on more specific datasets purpose built for a narrow application, such as speaking in the language of a medical doctor, or mimicking the manner and style of their individual user. In the near future, corporate AIs will be routinely fed your data, probably without your awareness or your consent. Any trustworthy AI system should transparently allow users to control what data it uses.

Many of us would welcome an AI-assisted writing application fine tuned with knowledge of which edits we have accepted in the past and which we did not. We would be more skeptical of a chatbot knowledgeable about which of their search results led to purchases and which did not.

You should also be informed of what an AI system can do on your behalf. Can it access other apps on your phone, and the data stored with them? Can it retrieve information from external sources, mixing your inputs with details from other places you may or may not trust? Can it send a message in your name (hopefully based on your input)? Weighing these types of risks and benefits will become an inherent part of our daily lives as AI-assistive tools become integrated with everything we do.

Realistically, we should all be preparing for a world where AI is not trustworthy. Because AI tools can be so incredibly useful, they will increasingly pervade our lives, whether we trust them or not. Being a digital citizen of the next quarter of the twenty-first century will require learning the basic ins and outs of LLMs so that you can assess their risks and limitations for a given use case. This will better prepare you to take advantage of AI tools, rather than be taken advantage by them.

In the world’s first few months of widespread use of models like ChatGPT, we’ve learned a lot about how AI creates risks for users. Everyone has heard by now that LLMs “hallucinate,” meaning that they make up “facts” in their outputs, because their predictive text generation systems are not constrained to fact check their own emanations. Many users learned in March that information they submit as prompts to systems like ChatGPT may not be kept private after a bug revealed users’ chats. Your chat histories are stored in systems that may be insecure.

Researchers have found numerous clever ways to trick chatbots into breaking their safety controls; these work largely because many of the “rules” applied to these systems are soft, like instructions given to a person, rather than hard, like coded limitations on a product’s functions. It’s as if we are trying to keep AI safe by asking it nicely to drive carefully, a hopeful instruction, rather than taking away its keys and placing definite constraints on its abilities.

These risks will grow as companies grant chatbot systems more capabilities. OpenAI is providing developers wide access to build tools on top of GPT: tools that give their AI systems access to your email, to your personal account information on websites, and to computer code. While OpenAI is applying safety protocols to these integrations, it’s not hard to imagine those being relaxed in a drive to make the tools more useful. It seems likewise inevitable that other companies will come along with less bashful strategies for securing AI market share.

Just like with any human, building trust with an AI will be hard won through interaction over time. We will need to test these systems in different contexts, observe their behavior, and build a mental model for how they will respond to our actions. Building trust in that way is only possible if these systems are transparent about their capabilities, what inputs they use and when they will share them, and whose interests they are evolving to represent.

This essay was written with Nathan Sanders, and previously appeared on Gizmodo.com.

Posted on May 11, 2023 at 7:17 AMView Comments

Sidebar photo of Bruce Schneier by Joe MacInnis.