Reported data breaches increased by 11% in the first half of 2025, putting the year on course to be a record-breaker. With 72 million victims, the breach at student information system PowerSchool was the most damaging, but the overall number of victims has actually declined. No wonder 74% of businesses are planning to improve their cybersecurity, according to a report from Beazley.
The British Standards Institution has issued the first global standard for AI audits in a bid to tame the “wild west” of unregulated AI providers. The standard lays out criteria for AI certification rather than addressing the AI solutions themselves. The aim is to ensure both credibility and consistency across AI audit processes.
The EU’s Code of Practice, intended to outline how to comply with the imminent AI Act’s obligations, is a voluntary framework — so Meta has refused to sign it. Meta claims the Code goes beyond what the AI Act actually provides, as well as creating uncertainties for developers. A range of tech companies, especially those in the AI race, have opposed the legislation and asked for the rollout to be delayed; EU says no.
Brands don’t want to send too many messages, so marketers have to choose which messages to send. CDP Hightouch’s new Smart Suppression feature makes this decision by predicting incremental lift from each message. Marketers can then suppress less effective messages based on percentile thresholds or probabilistic filtering.