AI offers faster R&D, but skills gap and data bias are barriers

The Pistoia Alliance, a global, not-for-profit alliance that works to lower barriers to innovation in life science and healthcare R&D, has today announced the results of a survey of life science professionals, on the implementation of AI and blockchain in the life sciences industry. The survey shows there is a high level of interest in AI among respondents, with 57 percent already engaging in computational drug repurposing. Similarly, the findings revealed that understanding of blockchain has increased, with 89 percent now aware of the technology, compared to 82 percent in 2018. Despite this increase, the survey identified that once again, lack of access to people with relevant blockchain skills remains the biggest barrier to widespread adoption (selected by 30 percent).

“The industry clearly has a willingness to engage with blockchain and AI technologies, but historical barriers are hampering progress. Cross-industry collaboration will be essential to overcoming issues around access to data and skills, so that more companies and thus, patients, can benefit from these technologies,” commented Dr Steve Arlington, President of the Pistoia Alliance. “70 percent of our survey participants think blockchain has the potential to make a real difference in patient data management and sharing. Blockchain’s ability to instantly create tamper-proof records will become a key part of increasing patient participation as more clinical trials are conducted remotely because of the pandemic. We hope the security advantages can both improve patient trust and facilitate further knowledge sharing across the life science community.”

Another recurring challenge identified in the survey was data quality and data standards. Behind skills, participants ranked lack of standards (19 percent) and interoperability (17 percent) among the next biggest barriers slowing blockchain adoption. Likewise, 38 percent think algorithmic bias poses a barrier to AI for drug repurposing, and a further 42 percent think it has potential to be a barrier. Life sciences generates huge volumes of data in an increasing number of formats. When data is disorganized and siloed it is not machine readable, and when information ‘training’ an algorithm is limited it eventually creates bias in the AI’s outputs. Organizations can address these data quality issues by adhering to the FAIR principles of Findable, Accessible, Interoperable and Reusable. The Pistoia Alliance has published a freely available toolkit to assist with FAIR implementation.

“Technologies including AI and blockchain have the potential to transform drug development. Yet no matter how powerful these technologies become, challenges and bias will exist until we improve the quality of data feeding algorithms,” commented Pistoia Alliance consultant Becky Upton. “To eliminate bias, data sets must be varied and drawn from accurate, diverse sources. Standards for data storing and sharing must also be improved. The Pistoia Alliance has created a Center of Excellence in AI and a project dedicated to Informed Consent using blockchain – to provide a space for the industry to share best practices and discuss common challenges. We urge any interested parties to get involved with our work and help inform our outputs, so that we can collectively continue to accelerate R&D.”