An artificially intelligent chat agent that answers adolescents' questions related to sex, drugs, and alcohol: an exploratory study.

PURPOSE: The aim of this study was to investigate if and how an artificially intelligent chat agent (chatbot) that answers questions about sex, drugs, and alcohol is used and evaluated by adolescents, especially in comparison with information lines and search engines. METHODS: A sample of 929 adolescents (64% girls, mean age = 15), varying in urbanization level and educational level, participated in this study. Use of the chatbot was objectively tracked through server registrations (e.g., frequency and duration of conversations with the chatbot, the number and topics of queries), and a web-based questionnaire was used to evaluate the chatbot (e.g., the perception of anonymity, conciseness, ease of use, fun, quality and quantity of information, and speed) and to compare it with information lines and search engines. RESULTS: The chatbot reached high school attendees in general and not only adolescents with previous experience related to sex, drugs, or alcohol; this is promising from an informed decision-making point of view. Frequency (M = 11) and duration of conversations (3:57 minutes) was high and the chatbot was evaluated positively, especially in comparison with information lines and search engines. CONCLUSION: The use of chatbots within the field of health promotion has a large potential to reach a varied group of adolescents and to provide them with answers to their questions related to sex, drugs, and alcohol.

Roughness scaling in cyclical surface growth.

The scaling behavior of cyclical growth (e.g., cycles of alternating deposition and desorption primary processes) is investigated theoretically and probed experimentally. The scaling approach to kinetic roughening is generalized to cyclical processes by substituting the number of cycles n for the time. The roughness is predicted to grow as n(beta) where beta is the cyclical growth exponent. The roughness saturates to a value that scales with the system size L as L(alpha), where alpha is the cyclical roughness exponent. The relations between the cyclical exponents and the corresponding exponents of the primary processes are studied. Exact relations are found for cycles composed of primary linear processes. An approximate renormalization group approach is introduced to analyze nonlinear effects in the primary processes. The analytical results are backed by extensive numerical simulations of different pairs of primary processes, both linear and nonlinear. Experimentally, silver surfaces are grown by a cyclical process composed of electrodeposition followed by 50% electrodissolution. The roughness is found to increase as a power law of n, consistent with the scaling behavior anticipated theoretically. Potential applications of cyclical scaling include accelerated testing of rechargeable batteries and improved chemotherapeutic treatment of cancerous tumors.

Scaling behavior of cyclical surface growth

The scaling behavior of cyclical surface growth (e.g., deposition/desorption), with the number of cycles, n, is investigated. The roughness of surfaces grown by two linear primary processes follows a scaling behavior with asymptotic exponents inherited from the dominant process while the effective amplitudes are determined by both. Relevant nonlinear effects in the primary processes may remain so or be rendered irrelevant. Numerical simulations for several pairs of generic primary processes confirm these conclusions. Experimental results for the surface roughness during cyclical electrodeposition/dissolution of silver show a power-law dependence on n, consistent with the scaling description.