AI can be considered a branch of psychology. Many psychologists contributed to its advancement over the past 70 years. More recently it has become a domain for engineers, computer scientists and data scientists. The basis of AI is that by using techniques inspired by neural biology (artificial neural network) computers can be enabled to learn in much the same way as organisms with biolological neural network are able to do. This approach has been very succesful, with AI now permeating many aspects of our lives, society and the economy.

A Psychologist learns to conduct psychological assessment and intervention by using neural networks in their brain. Now that we have machine learning by artificial neural networks: is it possible to develop robot machines that can deliver psychological services?

CogAI is a robot that performs psychological assessment (cognitive and functional) of an older person (60 years+), automatically over the phone. From the assessment results, CogAI can accurately (> 95% correct) conclude, for most people, that the person will not have an onset of dementia in the next 5 years.

This presentation will describe the development of CogAI and present results from studies of CogAI's psychometric properties. Briefly, CogAI was designed as a Conversational AI Chatbot, that asks assessment questions, listens to answers and scores responses. On completion of assessment items administration, CogAI tallies item scores and compares the results to normative data to draw clinical conclusions and make recommendations. Administration details, results, conclusions and recommendations are compiled into a Psychological Report, which then sent to the referring clinician.

Our studies have found that the CogAI agrees with the results of the same assessment administered, over the phone, by a psychologist (r=0.81), that the one week test-retest reliability of the CogAI (r=0.76) is good and that CogAI scores can be used to replicate study findings of psychologist administered assessments.

We conclude that the CogAI is a drop-in AI-automated replacement for the same or similar assessment by a psychologist.

AI has been shown (in our studies and others) to be able to automate specific units of psychological services provision. We need to start thinking about the pathway(s) forward for AI in Psychology.

The future will be a mix of augmentation (service by psychologist using AI), automation (service by AI alone) and innovation (services by AI that have yet to be imagined). Informed participation by Psychologists is the best way to shape that mix.

History shows that over time health outcomes for many physical ailments have improved dramatically, partially attributable to technological advances in medicine. Precision in measuring physiological characteristics such as, heart rate and temperature in the 19th century and blood pathology and MRIs in the 20th century are today mainstays of medical practice. Parallel advances in methods of measurement in psychology have been slower, as has the lack of dramatic improvement in mental health outcomes.

Among the mental health workforce psychologists are the only group with advanced training in psychometrics, which creates an opportunity to lead a measurement-based care revolution. Measurement-based care refers to two processes: routine assessments, such as measuring the severity of symptoms with rating scales, and the use of assessments in decision-making.

This presentation will review the state of measurement in psychology practice and conclude that the field is undergoing a revolution not seen since IQ assessments were developed in 1912. Just like the stethoscope and thermometer revolutionised medicine, advances in transtheoretical mental health assessment powered by artificial intelligence (AI) will propel psychology forward. These advances are not without risks, so a proactive and empowering blueprint of a possible future for psychologists will be presented.