#nasdaq-aapl — Public Fediverse posts

Apple Should Lead on AI Ethics, and This Shareholder Proposal Is a Good First Step

Apple shareholders will vote next month on a proposal brought by the AFL-CIO asking the company to report on its use of artificial intelligence and what ethical guidelines it has adopted around the use of AI.

The adoption of AI technology into business raises a number of significant social policy issues. For example, the use of AI in human resources decisions may raise concerns about discrimination or bias against employees. The use of AI to automate jobs may result in mass layoffs and the closing of entire facilities. AI may be used in ways that violate the privacy of customers and members of the public. AI technology may be used to generate “deep fake” media content that may result in the dissemination of false information in political elections.

The list could go on: the shareholders’ Supporting Statement could include concerns about AI’s carbon footprint or its military applications, for instance. But there’s enough here to reinforce a point I made in a recent post about the inclusion of Apple and other so-called Magnificent Seven stocks in ESG portfolios: their use of AI hasn’t gotten the scrutiny it deserves, and the outsize returns of big AI plays have overshadowed their risks.

Some of those risks may be as overblown as the current hype around AI, but it is incumbent on the Board to show investors that they are. Others are already the subject of litigation and the cause of social unrest:

We believe that the adoption of ethical guidelines for the use of AI can help improve our company’s bottom line by avoiding costly labor disruptions. In 2023, writers and performers went on strike against the Alliance of Motion Picture and Television Producers in part over concerns that the use of AI technology to create media content will infringe on the intellectual property and publicity rights of writers and performers and potentially displace human creators….

In our view, AI systems should not be trained on copyrighted works, or the voices, likenesses and performances of professional performers, without transparency, consent and compensation to creators and rights holders. We also believe that AI should not be used to create literary material, to replace or supplant the creative work of professional writers.

The example is well-chosen for a couple of reasons. First, Apple is a leading member of the Alliance of Motion Picture and Television Producers, and had to put deals on hold last year as the writers and actors’ strikes played out. Second, issues around copyright and IP can no longer be ignored. AI developers are facing lawsuits by the New York Times, Getty Images, and the Authors Guild, among others, for training their machines on copyrighted material.

In response, the developers have made noises about national security and American economic competitiveness, but the heart of their argument, as Michael Hiltzik sums it up, is that they can’t turn a profit, or the kinds of profits they’ve been turning, if they have to pay for copyrighted material or respect intellectual property rights. That’s the widely ridiculed Andreessen-Horowitz position. The leaders of the brave new world want a free pass to infringe the rights of others and violate copyright law, for the greater good, or bigger profits, whichever comes first. (Guess which comes first.)

Obviously, if courts and lawmakers don’t quite see things that way, there’s a significant downside risk for investors, and Apple doesn’t offer a cogent response to this concern.

Instead, the Board trots out boilerplate about Apple’s commitments to transparency and responsible development of its products (did ChatGPT write this?). All that amounts to little more than “trust us.” The Statement in Opposition to Proposal 7 also argues that the AFL-CIO request is “overbroad” and could require Apple to disclose information that might harm its competitive position; but the proposal specifically says reporting on AI should “omit any information that is proprietary, privileged, or violative of contractual obligations.” So that argument looks pretty wobbly.

What about Apple’s contention that reporting on AI use and ethical guidelines would be “premature in this developing area” where “the landscape is just starting to emerge and regulators around the world are actively engaged in new rulemaking”? It sounds as if the Board will stand on principle only after it figures out what it can get away with. Clear ethical guidelines would help the company navigate a changing landscape and comply with new rules and regulations, even or especially if new rules and regulations should allow companies to use AI unethically.

Finally, it’s worth noting that the Board never explicitly addresses the issue of labor disruptions highlighted in the AFL-CIO proposal. Its Statement in Opposition does not even once use the word “labor,” though I suppose that’s covered by the blanket statement that “existing guidelines, policies, and procedures already address the social issues raised.”

Apple tried but failed to exclude this proposal; the SEC denied its request at the start of the new year. Now that the Apple Board seems to have dug in its heels, shareholders should ask whether the Board is looking at the issue the right way. Putting together a report on AI use and ethics, and updating that report periodically, might open up a can of worms, if it’s not done with appropriate care, of course. But it could also open up new domains of research at Apple, set an example for other companies to follow, and even help temper and guide public debate over AI. Investors should be disappointed to see Apple relinquish leadership in this area and should vote for Proposal 7, if only to encourage the Board to think different.

Update 28 Feb 2024: The AFL-CIO proposal was rejected at today’s Apple shareholder meeting. But (29 Feb 2024) the proposal received 37.5% of the vote. That sends a clear signal the board cannot ignore. A good showing for a first outing.