MIT Technology Review on AI’s promise, harms and uncertainty

MIT Technology Review on AI’s promise, harms and uncertainty

MIT Technology Review published this piece on June 9, 2026, presenting a speaker’s account of major themes in AI. The speaker says generative AI tools are used for everyday office tasks, but there is almost no data yet on their effects on jobs and the economy. Knowing what happens inside companies, the speaker argues, is part of understanding what will happen to work. The piece contrasts distant fears of civilization-ending AI with harms it says have occurred: deepfakes used to incite violence, swing votes and sow distrust; lawsuits alleging chatbots encouraged or aided self-harm; and AI advice in warfare, where hurried decisions may leave little room to review outputs. On science, the speaker sees potential alongside warnings that AI may narrow research choices or contribute inaccurate results. The piece’s conclusion resists treating AI’s direction as inevitable: technologies comparable to electricity or the internet take time to settle and bring lasting change. The effects on jobs and the economy, and where AI is headed, remain unknown.

Publication: a9db2799-8abe-df0e-1cfb-4b3ff383d87d

attribution — uncertain: The piece is one first-person article, not a listing. The text never names its author, and the limitation discloses this. However, the title credits MIT Technology Review as the voice, while the arguments are one writer's personal view from an SXSW London talk. Calling that writer only 'the speaker' leaves the source's identity ambiguous for readers. Smallest repair: attribute to 'an MIT Technology Review writer, in a column based on an SXSW London talk.'

At SXSW London last week I gave a talk called “Five things you need to know about AI,” in which I shared what I think are the biggest themes in AI right now.

summary — uncertain: The reading drops one of the five themes: section 3, 'A lot of people really hate AI'. That section covers organized anti-AI protests, the data-center backlash over energy and electricity bills, QuitGPT, and violence. It also softens the author's voice. The tongue-in-cheek opener is lost, and 'distant fears' replaces the author's dismissal of doom scenarios as 'dystopian science fiction'. Smallest repair: add one sentence on the backlash and keep the author's characterization of doomer scenarios. Central terms such as deepfakes and generative AI are not explained, and there are no definitions.

Protests are getting more organized and drawing larger crowds.

date — pass: June 9, 2026 matches the UTC day of originalPublished. The source is a publisher page, so no archive disclosure is needed. The page was edited later to add videos, but the reading relies only on body text that is plainly original. The reading does not use the later related-story sidebar. Nothing is framed with hindsight, and OpenAI's 2028 goal is not presented as an outcome.

support — pass: Each claim matches the body text, and its excerpt is an exact quote. The claims stay attributed to the speaker. The note that warfare decisions may leave little room to review outputs follows the source's point about corner-cutting in active conflict. The question follows from the source's point that companies are still working out their use of AI. The 'some scientists warn' wording slightly merges two concerns, but stays within the source. The later edit added only videos and does not affect the cited text.

In fast-paced, high-stress active conflict, the risk that corners get cut is high.

media — pass: No media is selected, and there are no verified media candidates. The embedded talk videos are not available as checked candidates, so leaving media empty invents nothing.