- Pearson's AI improves student outcomes, unlike general AI models
- Pearson's shares have fallen 30% over AI replacement concerns
- Reports 6% rise in 2025 operating profit to 614 million pounds
LONDON, Feb 27 (Reuters) - British education company Pearson said its use of AI in its education courses was delivering better outcomes for students, in contrast to general AI models, which the company said had a negative impact on learning and reasoning skills.
Shares in Pearson, which provides assessments, virtual schools, textbooks and digital courseware, have fallen by around 30% in the last 12 months on concerns that AI could replace some of its products and services.
The shares were trading down 0.5% at 955 pence on Friday after it published full-year results.
Chief Executive Omar Abbosh said Pearson was scaling AI across its products, which were designed with teachers and had to meet exacting standards.
He said 80% of 2025's operating profit of 614 million pounds ($830 million) came from assessments and virtual schools.
"That requires a level of operational excellence that the accreditation authorities demand, and that you can prove works, sometimes in a court of law and often in a regulatory audit," he said, adding that it was "very defensible and robust".
Around 10% of profit came from digital courseware, which some investors viewed as more vulnerable to AI.
But Abbosh said general AI models did not compare to the skills Pearson was providing.
"We can show that learners learn better when they use our AI, which is not true when you use AI in the wild, actually that can cause cognitive atrophy problems," he said in an interview.
"In educational systems, the nearer you are to students and teachers the more trust that you need to have to operate. Pearson operates at a very high level of trust, and that's why I feel good about our business model."
Pearson reported a 6% rise in 2025 operating profit to 614 million pounds ($830 million) on sales of 3.58 billion pounds.
It said it expected profit to increase to 640 to 685 million pounds this year.
($1 = 0.7401 pounds)
Reporting by Paul Sandle; Editing by Susan Fenton and Jane Merriman
Source: Reuters