New tech on EFSA’s horizon: Crowdsourcing and cyber assistants

New technology like crowdsourcing and artificial intelligence computers is the future of risk assessment, the head of the European Food Safety Authority (EFSA) has said.

Speakers from various projects including IBM’s AI computer system 'Watson' and the UK crowdsourcing platform InnoCentive were invited to present at the authority’s second scientific conference at the Milan Expo last week.

Talking with us at the event, EFSA executive director Bernhard Url said he could envisage EFSA using a crowdsourcing ‘forum’ for its risk assessments in the future. 

“I think it’s a fascinating new tool for science to broaden the evidence base, to get input from another ‘crowd’.”

While there are no immediate plans to start crowdsourcing – he said it would likely begin with some small-scale, pilot projects and a selected ‘crowd’ rather than the general public.

url.jpg
Bernhard Url has been EFSA's executive director since last summer

“From my point of view we will not go out to the ‘crowd’, to the overall public, in the first step. But I could imagine that if we had a difficult scientific question that for example has to do with cardiology, why couldn’t we approach the European society of cardiology and ask them a question and distribute this via their networks?

"We would have access to thousands of scientists that usually don’t work with EFSA and for them it would be a question of two hours or maybe even 30 minutes to make a comment on that.”

He said this would tap into a “‘black sea’ of unknown expertise”.

Shifting the question of who can trust who

One attendee of the talks commented that the issue of an open EFSA had been raised originally due to questions on whether the public could trust scientists.

The arrival of crowdsourcing however had shifted this question to ask: Can scientists trust the public?

Url said this was something EFSA would have to consider both in terms of quality control and conflicts of interest.

However he said collective decision making by nature avoided individual’s potential bias from influencing conclusions. 

“In our panels process for example we have 21 experts in one panel, so even if one person had a hidden interest we were not aware of this person could not influence the decision making with 20 other colleagues. There is also a ‘safeguard’ with this collective decision making process.”

Pitching for business

Steven Drew, vice president of business development EMEA at the crowdsourcing platform InnoCentive, was one of the speakers at the event.

Within the InnoCentive platform organisations can set a ‘challenge’ with a cash reward for the best solution.

The crowd on that platform – researchers or business people for example – would then pitch in their ideas in the hope of winning the challenge.

He told us that of the top ten pharma firms in the world, he could think of only one InnoCentive hadn’t yet worked with.

Challenges ranged from market research questions to highly technical questions on new drugs.

On its website it lists NASA, Nature Publishing Group, Procter & Gamble,  Syngenta, The Economist, Thomson Reuters and government agencies in the US and Europe among its clients.

He said the system allowed organisations to be flexible with new demands on resources, essentially outsourcing when needed. 

Cyber assistants

Another speaker presented IMB’s artificial intelligence computer system Watson.

Cameron Brooks, director of public sector solutions for the Watson Group at IBM, said Watson could become an expert after six months to a year’s training – saving organisations time and resources. 

“What if there was a tool that had the time to read all the information available and was trained in your field?

“And you could ask it questions relating to you decisions like an assistant, a cyber assistant. That’s what Watson is endeavouring to be – a technology that can read 200 million pages, or 14 kilometres when it’s stacked up, and respond to a question in three seconds.” 

Brooks said this cognitive technology was the future and the risk of not investing in it far outweighed the risks of using it. 

He said it went beyond a simple internet search engine because Watson read and understood information in context. 

Watson has so far been ‘taken to medical school’ and is now being used to advise doctors in the US in the field of oncology.

“I’m excited about possibilities in a number of areas including food risk assessment,” he said.