AI needs clearer standards and transparency, not more police reporting

Officials pressed OpenAI after it declined to alert police about a user's flagged chat following the Tumbler Ridge attack; the author argues Canada should require clearer transparency and consistent standards from AI companies rather than lowering thresholds for reporting to law enforcement.

AI companies and government officials met after OpenAI flagged a user's chat connected to the Tumbler Ridge attack but did not notify police, saying the content did not meet its "imminent and credible risk" threshold. Ministers expressed disappointment and said the company must make changes or face possible regulation. The author warns that mandating broader reporting by intermediaries could undermine privacy and expand corporate surveillance. He argues that greater transparency about safety policies and uniform standards would better balance safety and privacy. Key details: - AI Minister Evan Solomon summoned OpenAI executives to explain why the company did not alert police. - OpenAI concluded the flagged account did not meet its standard of an "imminent and credible risk of serious physical harm to others." - Ministers publicly expressed disappointment and indicated government intervention if companies do not change practices. - The author calls for companies to disclose safety and escalation policies and for public transparency reporting on disclosures to authorities. Summary: The incident has intensified debate over when intermediaries should notify law enforcement and whether regulation should lower reporting thresholds. Ottawa has warned it may legislate if companies do not alter practices, while the article urges transparency and consistent standards as the preferred starting point. Undetermined at this time.