The chief executive of OpenAI has issued a rare personal apology after acknowledging his company failed to alert law enforcement about a banned user who went on to kill eight people in a school shooting in Canada. Sam Altman, in a letter posted on Friday to the community of Tumbler Ridge, British Columbia, expressed deep regret that ChatGPT usage patterns linked to 18-year-old Jesse Van Rootselaar were not reported to police before the February attack. The shooter, who died by suicide, had been barred from the platform the previous June for policy violations, but the company determined at the time the activity did not meet its internal threshold for a referral – a judgment that now carries tragic weight.

Viewed from British Columbia, the apology – addressed to Premier David Eby and published on local news sites – landed amid raw grief. Altman pledged to help prevent a recurrence, yet the episode has sharpened a broader tension between Silicon Valley’s content moderation paradigms and the public safety duties that governments increasingly expect. In Ottawa, the incident has revived calls for mandatory reporting requirements on artifical intelligence firms, akin to obligations long imposed on banks and telecoms, as lawmakers question whether voluntary standards are adequate when lives are at stake.

A continent away, a different legal reckoning is advancing in a California courtroom. Elon Musk, the billionaire co-founder who acrimoniously parted ways with OpenAI, has dramatically streamlined his lawsuit on the eve of trial. According to court filings, Musk abandoned all but two of his original 26 claims, dropping fraud and constructive fraud allegations to concentrate on his central narrative: that Altman and key backer Microsoft breached the company’s founding mission by restructuring it into a for-profit entity in 2019, turning a philanthropic pledge into what Musk calls a ‘wealth machine’. Jury selection is set to begin on Monday in Oakland, with opening arguments expected on Tuesday. Musk is seeking damages, and the trimmed-down case, though leaner, still threatens to expose internal tensions over a potential initial public offering that could value OpenAI at $1 trillion.

From Moscow, where business media followed the legal manoeuvring, the jettisoning of most charges was interpreted as a tactical retreat to keep jurors focused on a simple, emotive claim of betrayed idealism. The slimmed suit may prove no less perilous for Altman and his colleagues if it forces into the open early debates about the company’s trajectory.

In a peculiar coda to a bruising week, an obscure manufacturing entrepreneur paid $100,000 to unlock a paywalled podcast interview featuring Altman and OpenAI president Greg Brockman. Journalist Ashlee Vance had offered to release the conversation from behind the paywall for a five-figure sum, and Jim Belosic, CEO of laser-cutting firm SendCutSend, promptly obliged. The now-public episode provides a rare unvarnished exchange, yet the fact that their words had to be liberated by a stranger underscores the uneasy transparency surrounding the company.

Collectively, the developments illuminate the colliding pressures confronting OpenAI. Altman’s belated apology signals an agonising recognition that internal protocols can have lethal real-world consequences, while the Musk trial threatens to put the company’s governance itself in the dock. For an enterprise racing toward one of history’s largest public listings, the capacity to reconcile breakneck growth with ethical stewardship is now being tested from a Canadian logging town and a San Francisco courthouse simultaneously.