Columbia University Libraries Journals
Not a member yet
6467 research outputs found
Sort by
Smart Contract Accountability Problems: Default Oracle Liability as the Solution: Leana Ter-Martirosyan
Full text linkSmart contracts have emerged as a transformative force in contract law, leveraging blockchain technology to automate transactions and reduce reliance on human intermediaries. However, their widespread adoption is hindered by significant legal challenges, particularly in determining liability for transaction failures. This Note examines the accountability problems inherent in smart contracts, focusing on the critical role of oracles—third-party entities that feed external data into blockchain-based agreements. While existing scholarship explores the theoretical foundations and potential applications of smart contracts, this Note shifts focus to liability allocation and proposes a novel framework: default oracle liability. Under this proposal, oracles bear primary responsibility for transaction errors arising from inaccurate data sourcing or validation failures. If oracles demonstrate that they functioned correctly, liability shifts to smart contract developers, who are responsible for ensuring secure and error-free code. By clarifying accountability, this framework incentivizes higher standards for data accuracy and software integrity, ultimately fostering a more reliable and legally-viable environment for smart contracts to operate
Stylistic (A)synchrony in Human-Machine Dyadic Interaction: Investigating Co-Adaptation Through the Lens of Communicative Naturalness
No full textNaturalness in interaction is generally perceived as a necessary foundation for ensuring spontaneous, fluent and comfortable exchange of information. Though naturalness is a given, more or less, in interpersonal (i.e., human-human) interaction, this is not always the case for human-machine interaction. While large language models like ChatGPT undergo continual improvement, they still tend to fall short of natural “humanlike” communication (Voss & Waring, 2024)
Co-Adaptation in Human–Machine Interaction: What We Learned and Where To Go Next: Conclusion to the Forum
No full textThis forum set out to explore the question: How does co-adaptation unfold in an ecosystem of learner–ChatGPT dyadic interaction? Working with one graduate-level EFL learner from a seven-week interactional dataset from RECIPE4U, the three contributions approached this question from complementary perspectives: communicative naturalness and stylistic synchrony; formal and topical alignment; and cognitive-psychological trajectories indexed by LIWC and grounded in interactional moves
Laboratories of Reproductive Justice: State Amendments and the Right to Paid Family Leave
No full textThis Note proposes legal strategies for recognizing paid family leave as a constitutional entitlement under state law. Long viewed as laboratories for democratic experimentation, states can play a central role in areas where federal protections remain limited. In the aftermath of Dobbs v. Jackson Women’s Health Organization, several states—including Maryland, Michigan, Missouri, Montana, Ohio, and Vermont—amended their constitutions to guarantee a right to “reproductive freedom.” This Note argues that these amendments create fertile ground for rights development and should be read to include an affirmative right to paid family leave. Drawing on a reproductive justice framework, which defines reproductive autonomy as the ability to have children, not have children, and parent children in safe and sustainable conditions, this Note analyzes the text of these amendments and the historical landscape of state positive-rights jurisprudence
Sharing the Algorithm: The Tax Solution to Generative AI
No full textTax policy offers a core tool for mitigating the sweeping public policy challenges of generative artificial intelligence (“AI”). Specifically, we propose a tax that would allow the public to own a share of AI itself, not just through future income tax liabilities or new excise taxes, but a proposed ownership structure that requires a one-time tax payment by generative AI firms in the form of equity.
Fractional public ownership of AI would directly address four of the key harms of AI that have been well-documented in a deep and still expanding literature. First, many types of AI were built through the unauthorized use of millions of copyrighted works, allegedly amounting to copyright infringement on an unprecedented scale. Sharing ownership of AI would compensate injured creators alongside the broader public whose data was nonconsensually harvested. Second, AI is expected to pose massive labor market disruptions, but shared ownership would allow displaced workers to benefit from the profits of the technology substituting for their labor. Third, greater public voice in the corporate governance of AI could lead to greater scrutiny and bolder interventions in the ways AI has been shown to reproduce and compound many existing forms of discrimination. Finally, sharing the ownership of AI through government’s principal tool for redistribution, taxation, directly addresses the rapid wealth concentration and monopolization already underway with AI developers. This proposal can also work in tandem with targeted regulation of AI and private law remedies addressing AI’s many harms.
Ultimately, the original contribution of this Article is to propose a unique in-kind tax payment structure that would require firms with ownership of AI to remit equity shares to the public. The Article describes multiple structures for this arrangement, drawing from existing models of fractional ownership used in private investment to serve as a paradigm for a partial public interest in AI. In total, this Article argues that many of the greatest concerns related to AI can be solved through sharing AI. And tax policy is the best tool to achieve this goal
Pretty Privilege vs. Ingroup Bias in Decision Making
No full textIn-group and attractiveness bias are well-established concepts in social psychology. This study examines the concurrent influence of these concepts on the decision-making process using the Minimal Group Paradigm. Confederates, individuals who appear to be participants, were used to simulate out-group members. Participants (n = 119, aged 20–30 years) answered a series of mathematics questions, followed by a response, agreeing or disagreeing with the participant’s answer, from a confederate. Participants were then asked to rate the attractiveness of the confederates. Results indicated that in-group bias significantly outweighed attractiveness bias. Participants changed answers more frequently when their group disagreed, regardless of the confederate's attractiveness. Results highlighted the robust effects of group membership on decision-making. Additional research is required to explore confounds within decision-making, such as individual differences and familiarity bias
Response Time to Detect Careless Responding and Its Relationship with and Prediction of Emotional Distress
No full textPeople experiencing emotional distress struggle with cognitive and motivational decline, which has been correlated with patterns of careless responding. Although several methods have been used to detect careless responses in emotionally distressed respondents, the response time has not been widely explored. The current study conducted secondary data analyses on a sample (N = 37,819) who completed the Depression Anxiety Stress Scale (DASS-42) in an online survey between 2017 and 2019. First, a response-time-based approach––a normative threshold method––was used to identify careless responding and examine its association with emotional distress using the DASS-42. Second, four machine learning models––decision tree (DT), random forest (RF), support vector machine (SVM), and naive Bayes (NB)––were trained on DASS-42 item responses and response times to predict emotional distress severity level. A significant correlation was found between the number of careless responses and subscale scores of anxiety and stress. In addition, Mann-Whitney U tests showed statistically significant differences between careless and careful responders in depression, anxiety, and stress. Regarding the machine learning models, SVM was found to be the best predictive model for classifying distressed people with an accuracy, sensitivity, and specificity exceeding 90%. Our results suggest that, in addition to survey responses, response time can identify careless responders and predict distressed responders
Securities Fraud and the Market for Individual Stocks: Sue S. Guan
Full text linkAs long as stock markets have existed, so too have those who invest for idiosyncratic reasons unrelated to achieving financial returns. Some investors may be motivated by personal utility, others seek to signal loyalty to corporations, some might see their investments as an expression of their faith, others chase the latest fads, and still others simply make uninformed investment choices. Yet until relatively recently, these forms of demand-driven investing have received little attention. Most commentary has either dismissed the phenomenon as noise or attempted to absorb it into existing models of fundamental value-based investing.
This Article counters that understanding. It argues that demand-driven investing can create a market for individual stocks that is distinct from noisy trading as well as from fundamental-value-driven trading. This is especially so as the voices of retail investors, social media influencers (“finfluencers”), and other non-traditional, values-driven investors in today’s capital markets have grown in volume and strength. It is increasingly difficult to ignore the investors who systematically choose companies to invest in based on demand-driven factors such as alignment on environmental, social and governance (ESG) issues, an influential investor’s commentary on a security, and preferences between cultural and political values—in addition to seeking financial returns.
Understanding the demand-driven component of investor decision-making as creating a market for individual stocks yields fresh insights for today’s stock market information ecosystem and the securities laws’ ability to respond to misinformation within that ecosystem. Specifically, this Article offers a framework to explain how demand-driven
investing motivates investors and affects price discovery. This then challenges the limits of existing defenses against misinformation in the securities markets, in particular the scope of Rule 10b-5. Solely protecting financial information results in an incomplete understanding of investor motivations. On the other hand, the idiosyncratic nature of demand-driven investing can make determinations of liability inconsistent and unpredictable. This Article explores the complex doctrinal and policy questions that are implicated